Area under ROC (AUC) is a metric which is widely used for measuring the classification performance for imbalanced data. It is of theoretical and practical interest to develop online learning algorithms that maximizes AUC for large-scale data. A specific challenge in developing online AUC maximization algorithm is that the learning objective function is usually defined over a pair of training ex...

Pattern classification was an important part of the RBF neural network application. When the electronic nose is concerned, in many cases it is difficult to obtain the entire representative sample; it requires frequent updating the sample libraries and re-training the electronic nose. In addition,the gas detected from the online environment is not always the known gas in the training samples. Th...

In many online learning scenarios the loss functions are not memoryless, but rather depend on history. Our first contribution is a complete characterization of sufficient and necessary conditions for learning with memory, accompanied with a novel algorithm for this framework that attains the optimal O( √ T )-regret. This improves previous online learning algorithms that guaranteed O(T ) regret ...

In order to improve the monitoring effect of English online teaching, this paper combines deep learning algorithm construct teaching system and conduct real-time supervision process. Furthermore, by altering original DCGAN, research seeks apply approach constructing a convolutional adversarial network tackle issue small-sample target recognition in given scene develop an appropriate model. Acco...

In this paper, we study convergence and e ciency of the batch estimator and natural gradient algorithm for blind deconvolution. First, the blind deconvolution problem is formulated in the framework of a semiparametric model, and a family of estimating functions is derived for blind deconvolution. To improve the learning e ciency of the online algorithm, explicit standardized estimating function...

While determining model complexity is an important problem in machine learning, many feature learning algorithms rely on cross-validation to choose an optimal number of features, which is usually challenging for online learning from a massive stream of data. In this paper, we propose an incremental feature learning algorithm to determine the optimal model complexity for large-scale, online data...

A novel online identification method is developed for nonlinear multi-input multi-output process modeling issue, which is based on kernel learning framework and named as online kernel learning (OKL) algorithm in this paper. This proposed approach can adaptively control its complexity and thus acquire controlled generalization ability. The OKL algorithm performs first a forward increasing for in...

TD(λ) is the core temporal-difference algorithm for learning general state-value functions (Sutton 1988, Singh & Sutton 1996). True online TD(λ) is an improved version incorporating dutch traces (van Seijen & Sutton 2014, van Seijen, Mahmood, Pilarski & Sutton 2015). Emphatic TD(λ) is another variant that includes an “emphasis algorithm” that makes it sound for off-policy learning (Sutton, Mahm...